Photoelectrochemical water splitting using TiO₂/α-Fe₂O₃ heterojunction films produced by chemical vapour deposition

Abstract

This study reports the enhanced photoelectrochemical (PEC) performance of TiO₂/α-Fe₂O₃ heterostructure films fabricated via a sequential aerosol-assisted chemical vapour deposition (AACVD) of hematite at 450 °C, followed by atmospheric pressure CVD (APCVD) of anatase TiO₂ with controlled thickness. Structural analyses (XRD, Raman, XPS) confirmed phase purity and oxidation states, while UV-vis spectroscopy revealed a narrowed bandgap and extended visible light absorption for the heterostructures compared to pristine films. The optimized TiO₂/α-Fe₂O₃ (8 min) photoanode achieved a photocurrent density of 1.75 mA cm⁻² at 1.23 V vs. RHE in 1.0 M NaOH under AM 1.5G illumination, representing a ∼150% improvement over pure α-Fe₂O₃. Incident-photon-to-current efficiency (IPCE) reached 7.47% at 420 nm, with enhanced performance sustained across the visible range. Transient absorption spectroscopy (TAS) revealed prolonged charge carrier lifetimes, indicating suppressed electron–hole recombination. The heterojunction design also improved stability, maintaining performance for over 16 h compared to 6.5 h for hematite alone. These synergistic effects including narrowed bandgap, efficient charge separation, and enhanced light harvesting highlight the novelty of combining AACVD and APCVD in fabricating TiO₂/α-Fe₂O₃ heterostructures as durable, high-performance photoanodes for scalable solar hydrogen generation.

---

1. Introduction

Photoelectrochemical water splitting is a promising approach for sustainable energy production.^1–5 Generating electricity from hydrogen is highly desirable since it serves as an eco-friendly energy carrier and helps reduce reliance on fossil fuels. However, conventional hydrogen production methods using oil, coal, and methane have detrimental environmental effects, especially regarding climate change, due to the significant carbon dioxide emissions they generate.^6–8 The application of photoanodes in water splitting has garnered significant interest since the work of Fujishima and Honda (1972).⁹ Since that time, numerous studies have focused on enhancing PEC efficiency and improving catalyst stability.

TiO₂ naturally occurs in three crystalline forms, anatase, rutile, and brookite, and possesses a wide band gap (>3.0 eV), which limits its photo-absorption to the UV range.^10–12 However, hematite α-Fe₂O₃ features a lower bandgap (2.0–2.2 eV), enabling enhanced visible light absorption. Additionally, it is relatively non-toxic, chemically stable, cost-effective, and possesses a suitable valence band edge position.^13–18 These combined attributes establish hematite as one of the most promising metal oxides for water splitting.

Nevertheless, hematite has some notable limitations, one of which is its conduction band edge being positioned below the reversible hydrogen potential.¹⁹ Additionally, the photogenerated holes in Fe₂O₃ exhibit limited efficiency in driving the water oxidation process, this is attributed to the short hole diffusion length in Fe₂O₃,²⁰ in contrast to the high penetration depth.²¹ Several studies have employed hematite as a photoanode for oxygen evolution in water splitting; however, these materials require an external electrical bias to generate hydrogen.

It is important to note that, despite its theoretical potential, the actual efficiency of hematite for water splitting remains negligible in practice when compared to theoretical calculations (16.8%),^19,22 this is due to high charge recombination, limited hole diffusion length, and low electrical conductivity.

To enhance the efficiency of Fe₂O₃ in water splitting, various strategies have been employed, including doping it with different elements or combining α-Fe₂O₃ with other metal oxide semiconductors like TiO₂, Ta₂O₅, and WO₃.^19,23–27 In fact, the exploration of Fe₂O₃ for water splitting dates back quite a while, with the first paper on its use being published by Hardee and Bard.²⁸ They discovered that when Fe₂O₃ was placed on a Ti and Pt substrate, the photocurrent under 500 nm illumination in KCl was 0.8 V versus the saturated calomel electrode (V_SCE). In the same year, another study by Quinn et al. reported that on Fe₂O₃ single crystals under 475 nm illumination in 0.5 M NaOH, the photocurrent density reached 1.57 mA cm⁻² versus the reversible hydrogen electrode (V_RHE).²⁹

However, as mentioned, the PEC performance of α-Fe₂O₃ can be enhanced through surface treatment and/or metal ion doping, which helps reduce electron–hole recombination³⁰ and increases donor density.³¹ Surface treatment also leads to accelerated solar water oxidation.^32–34 It was found that heterojunction systems enhance PEC performance by facilitating efficient charge separation and reducing recombination of photogenerated electron–hole pairs. This leads to improved light absorption and higher photocurrent, boosting overall PEC water splitting efficiency. The study reports on WO₃/TiO₂ core/shell heterojunction photoanodes with a dendritic TiO₂ shell that significantly enhances photoelectrochemical water splitting performance.³⁵ The dendritic TiO₂ increases surface area and active sites, while the heterojunction improves charge separation and reduces recombination. This architecture leads to higher photocurrent density, better light absorption, and improved stability compared to individual WO₃ or TiO₂ electrodes, demonstrating the important role of dendritic TiO₂ in optimizing PEC efficiency. Another study focuses on surface engineering of CuO–Cu₂O heterojunction thin films to enhance their photoelectrochemical water splitting performance.³⁶ By optimizing the interface and surface properties, the heterojunction facilitates better charge separation and transport. As a result, the films exhibit increased photocurrent and improved stability during water splitting. In addition, the paper reports the fabrication of WO₃/BiVO₄ heterojunction photoanodes with nanostructured WO₃ to boost photoelectrochemical performance.³⁷ The nanostructuring enhances light absorption, charge separation, and interfacial charge transfer. Consequently, the photoanodes achieve higher photocurrent density and improved efficiency in water splitting.

In this paper, we present, for the first time, α-Fe₂O₃ deposited by AACVD on an FTO substrate at 450 °C, followed by anatase TiO₂ deposited as a second layer by APCVD and annealed at 500 °C to remove carbon contamination. This TiO₂/α-Fe₂O₃ layered approach enhanced both the water splitting efficiency and stability compared to pure α-Fe₂O₃ films. Additionally, TiO₂ films deposited using this method demonstrated excellent water splitting performance under sunlight. A pure Fe₂O₃ film, deposited via AACVD, proved effective for water splitting (0.8 mA cm⁻² at 1.23 V vs. RHE) and remained stable for about 6.5 hours. A notable improvement in photocurrent density was observed for the heterojunction after coupling with TiO₂ (1.75 mA cm⁻² at 1.23 V vs. RHE), which increased the stability to 16 hours. Transient absorption spectroscopy (TAS) was used to investigate charge carrier recombination dynamics and carrier lifetimes of the TiO₂/α-Fe₂O₃ films. A correlation between PEC efficiency and carrier lifetimes measured by TAS was identified.

2. Experimental section

2.1. General procedure

All chemicals used in these experiments were provided by Sigma-Aldrich Chemical Co. and were used without further purification. Deposition was carried out on a 10 × 20 mm² fluorine-doped tin oxide (FTO) glass. Before use, the FTO substrates were meticulously cleaned with acetone (99%), isopropanol (99.9%), and distilled water, and then dried in air.

2.2. Deposition conditions

[0.23 g, 0.905 mmol] of iron(II) acetylacetonate Fe[CH₃COCH=C(O)CH₃]₂ was dissolved in 60 mL of ethyl acetate in a glass bubbler. Aerosol-assisted chemical vapor deposition (AACVD) was then employed to deposit α-Fe₂O₃ thin films onto FTO glass substrate. The mist was transferred into the reactor using nitrogen (BOC Ltd, 99.99%) at a flow rate of 1.4 L min⁻¹, and deposition was carried out for 40 minutes at 450 °C. Afterward, the reactor was allowed to cool to room temperature under nitrogen before the films were removed and annealed at 500 °C in nitrogen at 2 L min⁻¹ for 30 minutes to eliminate carbon contamination.

At this point, 40 mL of titanium(IV) isopropoxide [Ti(OCH(CH₃)₂)₄] was introduced into a glass bubbler. Atmospheric pressure chemical vapor deposition (APCVD) was used to deposit TiO₂ films on the FTO glass substrate, with α-Fe₂O₃ films forming as the second layer. Concurrently, titanium(IV) isopropoxide was heated in the bubbler to around 120 °C, and the vapor was transferred to the reactor via nitrogen (BOC Ltd, 99.99%) at a flow rate of 4 L min⁻¹. Deposition took place at 450 °C for TiO₂ and TiO₂/α-Fe₂O₃ films for different time intervals (4, 8, and 12 minutes) on the glass substrate, followed by annealing at 500 °C for 30 minutes (Fig. S1).

3. Photoelectrochemical performance (PEC)

PEC measurements were carried out using a 1 M NaOH electrolyte (Fluka, in ID water, pH = 13.6) in a three-electrode electrochemical cell under simulated sunlight (100 mW cm⁻², AM 1.5G). A platinum wire served as the counter electrode, and an Ag/AgCl electrode in saturated KCl was used as the reference. The photoelectrode was controlled by a potentiostat (Ivium Technology), with the reversible hydrogen electrode (RHE) potential given by the equation:

The potential was scanned from −1.0 to 1.23 V at a rate of 50 mV s⁻¹, and sunlight was simulated using an Ivium CompactStat (IVIUM Technologies) Lamp at an intensity of 100 mW cm⁻² (AM 1.5G). The illumination intensity was calibrated using a silicon reference cell and optical meter (Newport, Model 1918-R). Incident-photon-to-current efficiency (IPCE) measurements were performed with 1.0 M NaOH electrolyte and a 200 W Xe lamp at 1.23 V vs. RHE. IPCE values were calculated using the formula:

where I_SC is the current density at 1.23 V vs. RHE, λ is the wavelength of the incident monochromatic light, and P_in is the light intensity at a specific wavelength. Mott–Schottky measurements were conducted in the dark at a 1 kHz frequency, using the same electrolyte as for IPCE.

The Applied Bias Photon-to-Current Efficiency (ABPE) was calculated according to the following equation:

where J is the photocurrent density (mA cm⁻²) measured by the electrochemical workstation, V_b is the applied bias versus RHE (V), and P_light is the incident light intensity under AM 1.5G (100 mW cm⁻²) illumination.

4. Transient absorption spectroscopy (TAS)

Transient absorption measurements were conducted using a custom-built pump–probe setup. The monochromatic pump pulse was generated by a Q-switched laser system (InnoLAS, Picolo AOT MOPA, Picosecond Nd:YVO₄ Laser System), where the fundamental frequency was tripled to produce a 355 nm output, with the pulse energy adjusted to 300 μJ cm⁻² using a reflective neutral medium filter. The broadband probe pulse, spanning from 420 nm to 1100 nm, was generated through a supercontinuum process by focusing a 1300 nm seed pulse onto a 3 mm c-cut sapphire crystal. This seed pulse was produced by an optical parametric amplifier (Light Conversion, TOPAS Prime) after being pumped by a commercial Ti:sapphire amplifier (Coherent Legend Elite Duo, 4.5 mJ, 3 kHz, 100 fs). The delay time between the pump and probe pulses was controlled by an electronic delay generator (Stanford Research System DG535) with a jitter of approximately 100 ps. The spot sizes of the laser pulses were determined by transmissivity through a 75 nm precision laser pinhole, with 20% for the pump and 50% for the probe, resulting in the pump pulse being over three times larger than the probe pulse. In the experiments, the pump and probe pulses were overlapped on the front surface of the samples. All measurements were performed at room temperature, and the samples were kept in a nitrogen-filled chamber.

5. Results and discussion

5.1. X-ray diffraction (XRD) and Raman spectroscopy

Powder X-ray diffraction (PXRD) was employed to examine the TiO₂/α-Fe₂O₃, pure α-Fe₂O₃, and pure TiO₂ films (Fig. 1a). The analysis revealed that the iron oxide was in the α-Fe₂O₃ phase, with diffraction peaks corresponding to the (012), (104), (110), (113), (024), (116), and (214) planes at 2θ values of 24.2°, 33.1°, 35.5°, 40.8°, 49.4°, 54.1°, and 62.5°, respectively. These findings align well with the reference data for the α-Fe₂O₃ phase.^38,39 No peaks corresponding to other iron oxide phases were detected. The PXRD patterns for pure TiO₂ revealed a typical anatase phase, with diffraction peaks at (101), (112), (200), and (204) corresponding to 2θ values of 25.3°, 38.6°, 48.0°, 55.1°, and 62.8°, respectively.⁴⁰ The TiO₂/α-Fe₂O₃ films, with varying TiO₂ thicknesses, exhibited the anatase phase of TiO₂. However, for the films with a deposition time of 4, 8 and 12 minutes, a peak for α-Fe₂O₃ was observed, as the TiO₂ layer was thin enough to reveal the α-Fe₂O₃ peak. Fig. 1b presents the Raman spectra of TiO₂/α-Fe₂O₃, pure α-Fe₂O₃, and pure TiO₂ films. The Raman peaks for the Fe₂O₃ films appeared at 225 cm⁻¹ (A_1g), 245 cm⁻¹ (E_g), 292 cm⁻¹ (E_g), 410 cm⁻¹ (E_g), 502 cm⁻¹ (A_1g), 497 cm⁻¹ (E_g), and 612 cm⁻¹ (E_g). The TiO₂ films in the anatase phase exhibited peaks at 141.5 cm⁻¹ (E_g), 397 cm⁻¹ (B_1g), 514 cm⁻¹ (A_1g), and 635 cm⁻¹ (E_g). The TiO₂/α-Fe₂O₃ films displayed peaks corresponding to both TiO₂ in the anatase phase and α-Fe₂O₃. These Raman spectra align with the XRD results, confirming the presence of iron oxide in the hematite phase and TiO₂ in the anatase phase.

Fig. 1 (a) Shows the powder X-ray diffraction (PXRD) patterns of α-Fe₂O₃, TiO₂, and TiO₂/α-Fe₂O₃ layer films with varying TiO₂ thicknesses. (b) Displays the Raman spectra (recorded using a 633 nm laser wavelength) for all these films. The α-Fe₂O₃ film was deposited via AACVD at 450 °C on a glass substrate and annealed at 500 °C, while the pure TiO₂ and TiO₂ on α-Fe₂O₃ films were deposited using APCVD at 450 °C and annealed at 500 °C with different deposition durations (4, 8, and 12 minutes).

5.2. UV-vis spectroscopy

The optical properties of TiO₂/Fe₂O₃ films, as well as pure Fe₂O₃ and TiO₂ films, were examined using UV-vis spectroscopy. Fig. 2a presents the absorption spectra of all the films. Pure TiO₂ exhibits absorption around 380 nm, as expected, while pure Fe₂O₃ and TiO₂/Fe₂O₃ films with deposition times of 4 and 8 minutes absorb near 600 nm, with absorption extending into the visible spectrum. The absorption of the TiO₂/Fe₂O₃ film (12 minutes) shows a slight shift towards the visible region, around 650 nm. Fig. 2a also shows the bandgaps of these films, which were calculated using Tauc plots.⁴¹ The bandgaps of pure Fe₂O₃ and TiO₂ are approximately 1.95 eV and 3.2 eV, respectively. For the TiO₂/Fe₂O₃ films, the bandgaps are 2.1 eV, 2.1 eV, and 1.78 eV for deposition times of 4, 8, and 12 minutes, respectively.

Fig. 2 (a) UV-vis absorption spectra of pure α-Fe₂O₃, TiO₂, and TiO₂/α-Fe₂O₃ films. (b) Displays the absorption spectra for all films, where pure TiO₂ absorbs around 380 nm and pure α-Fe₂O₃ film absorbs near 600 nm. The bandgaps, calculated using Tauc plots, are also shown in (b).

Fig. 3 presents the UV-vis transmission spectra for all samples, displaying the characteristic interference fringes for TiO₂ and α-Fe₂O₃ based films. The transmission of TiO₂ films is approximately 72% for wavelengths between 380 and 780 nm, increasing to around 83% in the near-infrared region. In contrast, the transparency of hematite films deposited by AACVD is lower than that of TiO₂ films in the visible range, but fluctuates between 72% and 63% in the near-infrared. The transmission spectra of TiO₂/α-Fe₂O₃ films with a 12 minutes deposition time exhibit the lowest transparency in the visible range (500–760 nm), while the TiO₂/α-Fe₂O₃ films with a 4 minutes deposition time show a slight increase in transparency in the 700–2500 nm range compared to pure TiO₂ films.

Fig. 3 Transmission spectra of α-Fe₂O₃, TiO₂ and TiO₂/α-Fe₂O₃ films deposited on glass substrate.

5.3. X-ray photoelectron spectroscopy (XPS)

X-ray photoelectron spectroscopy (XPS) was employed to analyze the surface-level chemical states and valence band characteristics of all iron (Fe) and titanium (Ti) films. The 2p_2/3 peaks corresponding to high-spin Fe²⁺ and Fe³⁺ ions displayed a notably broader profile compared to those observed in low-spin Fe²⁺ or metallic Fe⁽⁰⁾.⁴² Gupta and Sen have demonstrated that the broadening arises from three factors:

(1) The presence of unpaired 3d electrons in the photoionized Fe cation.

(2) Spin–orbit coupling involving the 2p core hole state.

(3) The incorporation of crystal field effects and electrostatic interactions.^43,44

According to the study by Gupta and Sen, corroborated by additional research,^45,46 the Fe 2p_3/2 peaks were analyzed using multiplet structure calculations. It was determined that the Fe 2p_3/2 spectrum in α-Fe₂O₃ consists of four distinct components at 709.8, 710.7, 711.4, and 712.3 eV. Additionally, a satellite peak appears around 719 eV due to charge transfer processes and shake-up effects. Another peak, associated with surface structure, is observed at approximately 715 eV.⁴⁵

The Fe spectrum was fitted using a Gaussian–Lorentzian function, revealing Fe 2p binding energies characteristic of Fe³⁺. The 2p_3/2 peaks were centered at 709.2, 710.1, 710.6, 711.7, and 712.7 eV, with a pre-peak at 708.1 eV and a satellite peak at 719.1 eV. These results closely align with our samples, as illustrated in (Fig. 4a).

Fig. 4 Surface X-ray photoelectron spectra (XPS) of the Fe 2p region for: (a) α-Fe₂O₃, and (b) Ti, which remained in the +4 oxidation state across all films, with 2p_3/2 peaks centered around 458.3 eV.

The XPS analysis of Ti 2p in pure TiO₂ and the α-Fe₂O₃ series revealed peaks at binding energies of 458.9 eV and 464.6 eV, corresponding to Ti–O bonds of Ti⁴⁺ within the TiO₂ lattice (Fig. 4b). The 2p_3/2 peaks were more intense than the 2p_1/2 peaks due to the greater degeneracy of 2p_3/2 (four states) compared to 2p_1/2 (two states) in spin–orbit (j–j) coupling. Additionally, no Ti³⁺ peak was detected in any of the samples.

Fig. 5 presents the XPS spectra of the valence band (VB) region for Fe₂O₃, TiO₂, and TiO₂/α-Fe₂O₃ films with varying TiO₂ thicknesses. While all VB spectra exhibit a main band spanning from 0 to approximately 10 eV, the VB spectrum of the α-Fe₂O₃ film also features three distinct bands at around 2.3, 4.8, and 7.5 eV, aligning well with the characteristic VB structure of α-Fe₂O₃.⁴⁶ The valence band (VB) spectra of TiO₂ and TiO₂/α-Fe₂O₃ films, deposited at different durations (4, 8, and 12 minutes), exhibit two primary peaks at approximately 4.2 and 7.5 eV. A slight shift in binding energy is observed as the TiO₂ film thickness increases compared to the pure TiO₂ film. The most significant distinction among the VB spectra of α-Fe₂O₃, TiO₂, and TiO₂/α-Fe₂O₃ samples is their intensity. The VB intensity of α-Fe₂O₃ is notably higher than that of TiO₂ or TiO₂/α-Fe₂O₃. This is attributed to the presence of Fe 3d (t_2g and e_g orbital symmetry) and O 2p hybridized states, along with the O 2p valence state. Additionally, the two lowest-energy peaks, at 2.3 and 4.8 eV, correspond to excitations of the 3e_g and 2t_2g orbitals, respectively. A third peak, located around 7.5 eV, is associated with the 2e_g orbital.⁴⁶ Additionally, the O 2p peaks overlap with the 4d orbitals, with the O 2p state appearing around 7.0 eV. Furthermore, the O 2p σ states overlap with the e_g peaks.⁴⁷ Conversely, the valence band (VB) spectra of TiO₂ films feature e_g and t_2g states at approximately 7.5 eV and 4.2 eV, respectively, arising from O 2p orbitals.⁴⁸

Fig. 5 Valence band XPS spectra of pure Fe₂O₃, TiO₂, and TiO₂/Fe₂O₃ films with deposition times of 4, 8, and 12 minutes. The VB spectra reveal a decrease in the band onset as the TiO₂ thickness increases.

5.4. Scanning electron microscopy (SEM)

The morphology of Fe₂O₃, TiO₂, and TiO₂/α-Fe₂O₃ films was analyzed using SEM (Fig. 6). The α-Fe₂O₃ films exhibited approximately spherical particles with an average diameter of 100–200 nm (Fig. 6a), while the pure TiO₂ films featured densely packed particles ranging from 200 to 400 nm in diameter (Fig. 6b).

Fig. 6 SEM images of: (a) α-Fe₂O₃, (b) pure TiO₂, (c) TiO₂/α-Fe₂O₃ (4 min), (d) TiO₂/α-Fe₂O₃ (8 min), and (e) TiO₂/α-Fe₂O₃ (12 min) films grown by AACVD and APCVD on FTO at 450 °C and then annealed at 500 °C. The thicknesses are shown in side-view SEM images.

The morphology of the TiO₂/α-Fe₂O₃ films differed from that of the pure α-Fe₂O₃ and TiO₂ films. The TiO₂/α-Fe₂O₃ film with a deposition time of 4 minutes displayed particles with diameters ranging from 50 to 100 nm (Fig. 6c). The film deposited for 8 minutes had a similar morphology to pure TiO₂, consisting of densely packed particles approximately 500 nm wide (Fig. 6d). The TiO₂/α-Fe₂O₃ film with a deposition time of 12 minutes featured densely packed, needle-like particles, typically 200–400 nm in length (Fig. 6e).

Side-view micrographs revealed that the thickness of the pure α-Fe₂O₃ film on the substrate ranged from 368 to 422 nm (Fig. 6a), while the TiO₂ film was approximately 470 nm thick (Fig. 6b). The thicknesses of the TiO₂ layers in the TiO₂/α-Fe₂O₃ films with deposition times of 4, 8, and 12 minutes were around 438, 500, and 633 nm, respectively (Fig. 6c, d, and e).

5.5. Atomic force microscopy (AFM)

To further investigate the film morphology, AFM was employed to analyze the surface topography of all samples in this study (Fig. 7). AFM measurements were performed at a 10 × 10 μm scale. The resulting images reveal that the particles are similar, closely packed, pseudo-spherical grains. However, as shown in Fig. 6a, Fe₂O₃ appears to consist of smaller grains and has a flatter surface compared to the other samples. AFM images of TiO₂ and the TiO₂/α-Fe₂O₃ series showed particle aggregates, with varying surface roughness. This variation was quantified by measuring the RMS (root mean square) roughness of all samples at the 10 × 10 μm scale. The roughness of TiO₂ was 20.3 nm, nearly twice that of α-Fe₂O₃ (10.2 nm). The roughness values for the TiO₂/α-Fe₂O₃ (4 min) and TiO₂/α-Fe₂O₃ (12 min) samples were relatively similar, at 36.7 nm and 31.7 nm, respectively. In contrast, the TiO₂/α-Fe₂O₃ (8 min) sample exhibited a significantly lower surface roughness of about 19.8 nm. Additionally, the surface area of all samples was similar, with α-Fe₂O₃ having an area of 101 μm², and TiO₂, TiO₂/α-Fe₂O₃ (4 min, 8 min, and 12 min) having areas of 103, 104, 102, and 105 μm², respectively.

Fig. 7 2D (top) and 3D (bottom) surface images of: (a) α-Fe₂O₃, (b) TiO₂, and (c, d, and e) TiO₂/α-Fe₂O₃ films with deposition times of 4, 8, and 12 minutes, respectively, at a scale of 10 × 10 μm.

6. Photoelectrochemical (PEC) performances

Fig. 8 displays the current–voltage (I–V) curves for (a) TiO₂/α-Fe₂O₃ layers and pure α-Fe₂O₃, and (b) TiO₂ films, under both dark and simulated sunlight conditions (AM 1.5G, 100 mW cm⁻²). As seen in Fig. 8a, the photocurrent of hematite in the dark increases sharply at a voltage of approximately 1.70 V_RHE. Under simulated sunlight, the onset potential shifts to around 0.4 V_RHE, with the photocurrent rising until about 1.4 V_RHE. The Fe₂O₃ photocurrent peaks at approximately 0.7 mA cm⁻² at 1.23 V_RHE, where the RHE corresponds to the potential of the reversible oxygen electrode.

Fig. 8 PEC performance of (a) hematite and TiO₂/hematite films with varying deposition times (4, 8, and 12 minutes); (b) TiO₂ films. All films were deposited on FTO glass. Hematite was deposited using AACVD at 450 °C with Fe(II) acetylacetonate dissolved in ethyl acetate. TiO₂ films were prepared via APCVD using titanium isopropoxide at 450 °C, followed by a 1-hour annealing under nitrogen to eliminate carbon contamination. Water oxidation was tested in 1.0 M NaOH (pH 13.7) under both dark and AM 1.5G illumination (100 mW cm⁻²).

Fig. 8a also presents the photocurrent graphs for TiO₂/α-Fe₂O₃ films with varying TiO₂ thicknesses under sunlight conditions. It is evident that the TiO₂ layer, deposited by APCVD, not only enhances the photocurrent of α-Fe₂O₃ but also improves the onset potential. Specifically, the photocurrents of TiO₂/α-Fe₂O₃ (4 and 8 min) show onset potentials at 0.1 and 0.5 V_RHE, respectively, with currents increasing to approximately 1.2 and 1.75 mA cm⁻² at 1.23 V_RHE. The sample with an 8-minute deposition time achieves a photocurrent of around 2.3 mA cm⁻² at 1.45 V_RHE. However, for the thicker TiO₂ layer (12 min deposition time), the photocurrent decreased compared to pure hematite, reaching approximately 0.3 mA cm⁻² at 1.2 V_RHE. For all samples, the J–V curves in the dark conditions increased up to 1.67 V_RHE.

The highest photocurrent recorded for Fe₂O₃ to date was reported by Michael Grätzel et al. in 2011, reaching approximately 2.3 mA cm⁻² at 1.23 RHE.¹⁵ Guo et al.⁴⁹ enhanced the performance of the hematite photoanode by doping hematite nanorod films with Ru, resulting in the best performance to date, achieving approximately 5.7 mA cm⁻² at 1.23 RHE. The pure TiO₂ (with an 8-minute deposition time) was fabricated onto PEC photoanodes, and its performance was evaluated in 1.0 M NaOH under both dark and simulated sunlight (AM 1.5G, 100 mW cm⁻²) conditions. As shown in Fig. 8b, the photocurrent density of this TiO₂ sample under sunlight conditions was approximately 1.0 mA cm⁻² at 1.23 V_RHE, which is higher than previously reported for undoped TiO₂. Under dark conditions, the photocurrent of the same sample increased rapidly at around 1.9 V_RHE. However, an optimal TiO₂ thickness on α-Fe₂O₃ can enhance hole injection through surface contact with the solution, thereby reducing electron–hole recombination – a phenomenon confirmed by transient absorption spectroscopy (TAS) (see the TAS section). Although the optimized TiO₂/α-Fe₂O₃ heterojunction achieved a photocurrent density (1.75 mA cm⁻² at 1.23 V vs. RHE) slightly lower than the record value reported for hematite nanostructures, the present work demonstrates a scalable, low-temperature AACVD–APCVD fabrication route that delivers competitive efficiency, significantly enhanced operational stability, and mechanistic insight via TAS into charge carrier dynamics. These combined advances highlight the practical potential of TiO₂/α-Fe₂O₃ photoanodes for durable, large-area solar hydrogen generation. However, Table 1 provides a comparison of photocurrents reported for various Fe₂O₃ photoanodes in previous studies with our TiO₂/α-Fe₂O₃ photoanode.

Table 1 Comparing the photocurrents of some α-Fe₂O₃ heterojunction photoanodes obtained from the literature with our work

Photocathode	Electrolyte	Photocurrent density (mA cm^−2)	References
TiO2/Fe2O3 core/shell nanostructure	1 M KOH	0.91 mA cm^−2 at 1.23 V vs. RHE	50
Fe2O3/g-C3N4 heterojunction	1 M KOH	1.02 mA cm^−2 at 1.23 V vs. RHE	51
CQDs modified TiO2/Sn–Fe2O3−x heterojunction	1 M KOH	1.47 mA cm^−2 at 1.23 V vs. RHE	52
WO3/Fe2O3 nanocomposite	0.5 M Na2SO4	2.34 mA cm^−2 at 1.4 V vs. RHE	53
BiVO4/WO3/W heterojunction	0.1 M KH2PO4	2.01 mA cm^−1−2 at 0.6 V (vs. Ag/AgCl)	54
α-Fe2O3/CuO	0.05 M Na2SO4	0.53 mA cm^−2 at 1.0 V vs. RHE	55
ZnO/Fe2O3 core–shell nanowires	1 M NaOH	1.5 mA cm^−2 at 0.6 V vs. Ag/AgCl	56
TiO2/α-Fe2O3	1 M NaOH	1.7 mA cm^−2 at 1.23 V vs. RHE	This work

---

The positions of the valence and conduction band edges of TiO₂ and Fe₂O₃ films in contact with an electrolyte at pH 13.7, relative to the reversible hydrogen electrode (RHE) potential, were determined from the flat band measurements (Fig. 9). The band gaps of TiO₂ and Fe₂O₃ were found to be 3.2 eV and 1.95 eV, respectively. The conduction band of TiO₂ is positioned at 0.16 eV, while hematite shows a conduction band at 0.46 eV. Both of these conduction band edges are below the H₂O/H₂ electrochemical level for water reduction, meaning they cannot reduce H₂O without the application of an external bias (Fig. 9). Additionally, the conduction band of the Fe₂O₃ film is more positive than that of anatase TiO₂, which aligns with the photocurrent onset potential observed in Fig. 8. The valence band of the TiO₂ film lies significantly deeper than that of the Fe₂O₃ film and even falls below the water oxidation potential of 1.23 V vs. NHE, making it capable of oxidizing H₂O to O₂ (Fig. 9).

Fig. 9 Positions of the valence and conduction bands of TiO₂ and Fe₂O₃ thin films at pH = 13.7.

However, the heterogeneous system consisting of TiO₂ deposited on α-Fe₂O₃ thin film, which is itself on a FTO glass substrate, exhibits a type-II staggered band alignment in the 1.0 M NaOH electrolyte (pH 14). This specific alignment is critical for efficient charge separation and is depicted in the energy diagram. Upon illumination, both semiconductors absorb photons and generate electron–hole pairs. At the interface, the unique band alignment drives the photogenerated electrons from the higher-energy conduction band of Fe₂O₃ into the lower-energy conduction band of TiO₂. Simultaneously, holes from the higher-energy valence band of TiO₂ transfer to the lower-energy valence band of Fe₂O₃. This spatial separation prevents the recombination of electrons and holes. The electrons that accumulate in the TiO₂ conduction band travel through the external circuit to the Pt counter electrode to facilitate the hydrogen evolution reaction (HER). Meanwhile, the holes accumulated in the Fe₂O₃ valence band migrate to the semiconductor–electrolyte interface to participate in the oxygen evolution reaction (OER), completing the water-splitting process.^57,58

For all samples, the incident-photon-to-current efficiency (IPCE) at 1.23 V vs. RHE was measured as a function of various incident light wavelengths to compare the behavior of TiO₂ films with Fe₂O₃ and TiO₂/Fe₂O₃ films. The results are presented in Fig. 10. The TiO₂/Fe₂O₃ samples showed enhanced IPCE values compared to pure Fe₂O₃ across the wavelength range of 300–550 nm, especially in the visible region. For instance, the IPCE of TiO₂/Fe₂O₃ (8 min) at 420 nm increased to 7.47%, up from 3.9% for Fe₂O₃ alone. A similar improvement was observed for TiO₂/Fe₂O₃ (4 min), which reached 6.5%. In the UV region, IPCE values for TiO₂/Fe₂O₃ (4 min) were comparable to pure Fe₂O₃, while the values for TiO₂/Fe₂O₃ (12 min) decreased. Notably, at 500 nm, the IPCE of Fe₂O₃ dropped to zero, whereas TiO₂/Fe₂O₃ (8 min) maintained an IPCE value of 1.95% at this wavelength.

Fig. 10 Incident-photon-to-current efficiency (IPCE) of TiO₂, Fe₂O₃, and TiO₂/Fe₂O₃ films.

These results align with previous studies, which have reported a greater increase in IPCE values in the visible region compared to the UV region.⁵⁹ In contrast, studies on Zr, Sn, and Ti doping show that these elements have a minimal impact on the IPCE values of hematite in the visible region but a significant effect in the UV region.^25,30,60

Mott–Schottky (M–S) calculations were employed to determine the flat band potential (V_fb) and donor density (N_d) of Fe₂O₃, TiO₂, and TiO₂/Fe₂O₃ samples. Measurements were taken in the dark at a frequency of 1 kHz, using the equation:

where A is the active area, e is electron charge, ε₀ is the permittivity of a vacuum, ε is the dielectric constant, T is the absolute temperature, k is the Boltzmann constant and E is the applied potential.^61,62 The results are shown in Fig. 11.

Fig. 11 Mott–Schottky plots for three electrodes: (a) Fe₂O₃ and TiO₂/α-Fe₂O₃ films, and (b) anatase TiO₂ as the photoanode. The flat band potential and donor density of the samples were determined from this analysis.

All films exhibited a positive gradient, confirming that Fe₂O₃, TiO₂, and TiO₂/Fe₂O₃ are n-type semiconductors. As shown in Fig. 11a, the Fe₂O₃ films have a flat band potential around 0.48 V_RHE and a donor density of approximately 3.27 × 10¹⁹ cm⁻³. The TiO₂ thin films display a V_fb of about 0.16 V vs. RHE and a donor density of 1.68 × 10¹⁸ cm⁻³, as seen in Fig. 11b. Fig. 11a also shows that TiO₂/Fe₂O₃ films (4, 8, and 12 min) exhibit V_fb values of approximately 0.06, 0.52, and 0.74 V_RHE, with corresponding donor densities of 2.24 × 10¹⁹, 4.66 × 10¹⁹, and 1.75 × 10¹⁹ cm⁻³, respectively. The shift in the V_fb values for TiO₂/Fe₂O₃ films could be attributed to efficient hole transport from hematite to TiO₂, leading to enhanced water oxidation reactions at the surface.⁵⁹ However, since the V_fb of Fe₂O₃ typically ranges between 0.4 and 0.8 V_RHE, our results for Fe₂O₃ are generally consistent with those reported by Dotan et al.⁶³

The photoconversion efficiencies of the various photoanodes were assessed using the applied bias photon-to-current conversion efficiency (ABPE), derived from the LSV curves. As illustrated in Fig. 12a, the α-Fe₂O₃ photoanode exhibits a maximum ABPE of only 0.1% at 0.86 V versus RHE. The photoanodes prepared for 4 and 12 minutes show ABPE peaks of 0.19% (at 0.9 V vs. RHE) and 0.04% (at 1.0 V vs. RHE), respectively. Notably, the 8-minute photoanode displays a ABPE peak of 0.16% near 1.1 V vs. RHE, surpassing that of the bare α-Fe₂O₃ and occurring at a more practically relevant bias for PEC applications.

Fig. 12 (a) Applied bias photon-to-current efficiency (ABPE) curves for α-Fe₂O₃ and TiO₂/α-Fe₂O₃ with varying deposition times (4, 8, and 12 minutes). (b) Electrochemical impedance spectroscopy (EIS) for all samples in 1.0 M NaOH electrolyte.

However, the electrochemical impedance spectroscopy (EIS) measurements provide clear insight into the influence of coating duration on the interfacial charge transfer properties of the photoanodes. The Nyquist plots, which represent the imaginary and real components of the EIS plots (Z′ versus −Z′′), were measured at 1.23 V_RHE under simulated sunlight conditions (AM 1.5G, 100 mW cm⁻²) (Fig. 12b). The bare Fe₂O₃ electrode exhibits a semicircle with moderate diameter, reflecting relatively high charge transfer resistance (R_ct) due to its intrinsically poor conductivity and high recombination rates. Upon coating for 4 min, the semicircle diameter decreases slightly, suggesting reduced interfacial resistance and improved charge transfer kinetics, likely due to the formation of a thin layer that facilitates more efficient electron transport. The 8 min coated sample shows a comparable semicircle to Fe₂O₃, indicating only moderate improvement, as increased thickness can introduce additional scattering centers or recombination pathways. In contrast, the 12 min coated sample displays the largest semicircle, corresponding to the highest R_ct, which can be attributed to excessive coating thickness that hinders charge mobility, blocks active sites, and limits effective interaction with the electrolyte.

6.1. Long-term stability

The stability of photoelectrodes is crucial for the generation and commercial viability of PEC hydrogen production devices. As a result, enhancing the stability of semiconductors has been a focus of recent research. Fig. 13 illustrates the stability of hematite and TiO₂/α-Fe₂O₃ (8 min) films under simulated sunlight at an intensity of 100 mW cm⁻². The Fe₂O₃ film, deposited via AACVD, remained stable for over 6.5 hours, the photocurrent response measured at 1.0 V vs. RHE. Remarkably, the incorporation of a TiO₂ layer improved the stability of the α-Fe₂O₃ film, extending its performance to over 16 hours.

Fig. 13 Polarization curves of the Fe₂O₃ films (blue) and TiO₂/α-Fe₂O₃ films (red) recorded under simulated solar illumination (100 mW cm⁻²) with a fixed bias of 1 mA cm⁻². The measurements were conducted in a 1.0 M NaOH electrolyte under sunlight.

The morphology of the TiO₂/α-Fe₂O₃ (8 min) film was analyzed using SEM before and after PEC and stability tests. As shown in Fig. 14, the initial morphology featured densely packed particles approximately 500 nm in width (Fig. 14a). However, after undergoing prolonged testing (over 16 hours), the structure transformed into irregular shapes with visible pinholes, voids, and cracks (Fig. 14b). These morphological changes suggest that the sample's stability is influenced by the TiO₂ surface, as TiO₂ exhibits greater stability compared to Fe₂O₃.

Fig. 14 SEM images of the TiO₂/α-Fe₂O₃ (8 min) anode captured (a) before and (b) after stability tests. The photocurrent stability test was conducted for over 16 hours under simulated solar illumination (100 mW cm⁻²) with a constant bias of 1.0 mA cm⁻².

Fig. 15 presents the XPS analysis of the TiO₂/α-Fe₂O₃ (8 min) sample before and after stability tests. The results reveal a reduction in XPS intensity for Ti 2p and Fe 2p following the test, indicating that both TiO₂ and α-Fe₂O₃ underwent decomposition.

Fig. 15 X-ray photoelectron spectroscopy (XPS) analysis of the surface for (a) Ti 2p and (b) Fe 2p, conducted before and after the stability test.

7. Transient absorption spectroscopy (TAS)

Transient absorption spectroscopy (TAS) was employed to study the charge carrier dynamics in TiO₂, α-Fe₂O₃, and TiO₂/Fe₂O₃ films over a nanosecond to microsecond timescale at room temperature (25 °C). Previous studies have demonstrated that varying the thickness of TiO₂ on α-Fe₂O₃ enhances photocurrent generation in water-splitting reactions. In this work, TAS was utilized to examine how coupling hematite with TiO₂ influences the dynamics and lifetime of photo-generated charge carriers. Fig. 16 presents the transient absorption spectra of TiO₂, Fe₂O₃, and TiO₂/α-Fe₂O₃ samples. Pristine TiO₂ films exhibit a broad and structured photo-induced absorption (PIA) band with peaks at 630 and 770 nm, aligning with previously reported findings.⁶⁴ As reported by Akihiro Furube et al.,⁶⁴ the broad PIA band comprises multiple components: the 630 nm PIA band is primarily associated with trapped holes, while the 770 nm PIA band corresponds to trapped electrons. Our findings confirm that photogenerated hole absorption occurs around 630 nm, though other studies have suggested a spectral range of 450–570 nm with a lifetime spanning microseconds to milliseconds. However, this interpretation remains a topic of ongoing discussion.⁶⁴ The transient absorption data also reveal that the 770 nm PIA band disappears after 100 ns, whereas the 630 nm PIA band remains even beyond 1 μs. This indicates that the electrons have a shorter lifespan than the holes. The TA spectra of Fe₂O₃ sample exhibited an absorption peak at 560 nm, accompanied by a slight tail at 680 nm. This could be linked to trapped photo-generated holes and electrons, or potentially result from reactive oxygen species on the surface.^65–67 Moreover, the signal intensity of pure hematite is twice that of pure TiO₂, as shown in Fig. 16. The behavior of photo-generated holes in TiO₂ and Fe₂O₃ has been studied before using different hole scavengers, such as methanol for TiO₂, and thiocyanate, isopropanol, and iodide for Fe₂O₃.^67,68 The decay dynamics of photo-generated holes in Fe₂O₃ showed no dependence on the chemical environment when measured with TAS at 580 nm over a time scale of μs–ms. In contrast, the decay dynamics of photo-generated holes in TiO₂ exhibited a significant sensitivity to methanol at 460 nm.⁶⁸

Fig. 16 Transient absorption spectra at various delay times for pure TiO₂, pure Fe₂O₃, and TiO₂/Fe₂O₃ thin films. The pump wavelength is 355 nm, with a fluence density of 300 μJ cm⁻².

In our experiments, the TiO₂/Fe₂O₃ samples exhibit absorption at 560 and 700 nm, which we attribute to trapped photogenerated holes in Fe₂O₃, as the holes in Fe₂O₃ absorb significantly more strongly than those in TiO₂. This may cause a spectral overlap of trapped holes in Fe₂O₃ with those in TiO₂. However, when the TiO₂ thickness is increased (as in the 12-minute sample), absorption bands at 560 and 650 nm are observed for Fe₂O₃ and TiO₂, respectively, along with a band at 770 nm for electrons.

The TA spectra of the series of TiO₂/Fe₂O₃ films illustrate how the spectra change as the TiO₂ layer thickness increases (i.e., with longer reaction times). For the thinnest TiO₂ layer, the TA spectra show a larger contribution from Fe₂O₃, with a prominent PIA band peaking around 560 nm. As the TiO₂ layer thickness grows, the contribution from TiO₂ gradually increases, and the 650 nm PIA band becomes more prominent than the 560 nm PIA band.

The temporal evolution of the transient absorption decay was measured at 560 nm for TiO₂ excited at 355 nm with pulse energy densities of 300 μJ cm⁻² in the sub-nanosecond range. For TiO₂, the fastest TA decay occurred at 560 nm with a lifetime of t_50% = 85 ns, where t_50% represents the time for the signal to decay to 50% of its initial amplitude. For hematite, the TA decay at 560 nm showed an initial signal intensity of about 0.013 ΔO.D. with a lifetime of t_50% = 112 ns. The TiO₂/α-Fe₂O₃ (4, 8, 12 min) samples exhibited initial signal intensities of 0.075, 0.006, and 0.0045 ΔO.D. at 560 nm, respectively, with lifetimes of t_50% = 66 ns, 1200 ns, and 880 ns. Notably, the TiO₂/α-Fe₂O₃ (8 min) sample showed slower decays and a significant increase in signal intensity at 560 nm compared to pure Fe₂O₃ and TiO₂. These findings align with the photocurrent data obtained during water oxidation (Fig. 17), suggesting that the low efficiency of pure Fe₂O₃ is due to rapid electron–hole recombination. However, at 560 nm, single-layer TiO₂ shows a lower signal intensity than Fe₂O₃ because the valence band of Fe₂O₃ is higher than that of TiO₂. In contrast, TiO₂/Fe₂O₃ films, particularly the TiO₂/α-Fe₂O₃ (8 min) sample, exhibit much slower recombination, which could be attributed to the transfer of photogenerated holes from Fe₂O₃ to TiO₂. Additionally, photogenerated electrons in TiO₂ may be transferred to Fe₂O₃, extending the photo hole lifetime in the TiO₂ films.

Fig. 17 Transient absorption decay profiles at 560 nm for pure TiO₂, pure Fe₂O₃, and TiO₂/Fe₂O₃ thin films.

8. Conclusion

Fe₂O₃ film was deposited on FTO glass using aerosol-assisted chemical vapor deposition (AACVD) at 450 °C, employing iron(II) acetylacetonate [Fe(CH₃COCH=C(O)CH₃)₂] as the precursor. Subsequently, titanium isopropoxide [Ti(OCH(CH₃)₂)₄] was used in atmospheric pressure chemical vapor deposition (APCVD) to deposit TiO₂ on the Fe₂O₃ films, with deposition times of 4, 8, and 12 minutes. XRD and Raman spectroscopy confirmed the phase formation of the resulting samples, with Fe₂O₃ and TiO₂ present only in the hematite and anatase phases, respectively. XPS studies indicated that Fe and Ti were in the Fe³⁺ and Ti⁴⁺ oxidation states. The bandgap of TiO₂/Fe₂O₃ films was found to be smaller than that of Fe₂O₃ films, and a significant increase in photocurrent was observed in TiO₂/Fe₂O₃ films compared to pure Fe₂O₃ or TiO₂ films. PEC measurements of TiO₂/Fe₂O₃ films deposited on FTO showed a photocurrent of 1.7 mA cm⁻² at 1.23 V vs. NHE in 1.0 M NaOH, under AM 1.5 sunlight conditions. Transient absorption spectroscopy (TAS) revealed that TiO₂/Fe₂O₃ had a longer lifetime compared to pure hematite, which could be due to reduced electron–hole recombination, contributing to the improved photocurrent.

This study highlights the potential of TiO₂/Fe₂O₃ films as a promising material for water splitting with high stability.

Conflicts of interest

There are no conflicts to declare.

Data availability

Additional datasets available upon reasonable request.

The supporting data are available within the article and its SI. See DOI: https://doi.org/10.1039/d5ra05064b.

References

1. C. Bailleux, Int. J. Hydrogen Energy, 1981, 6, 461–471.
2. A. Landman, H. Dotan, G. E. Shter, M. Wullenkord, A. Houaijia, A. Maljusch, G. S. Grader and A. Rothschild, Nat. Mater., 2017, 16, 646–651.
3. C.-H. Liao, C.-W. Huang and J. C. S. Wu, Catalysts, 2012, 2, 490–516.
4. K. Mazloomi and C. Gomes, Renewable Sustainable Energy Rev., 2012, 16, 3024–3033.
5. S. J. A. Moniz, S. A. Shevlin, D. J. Martin, Z.-X. Guo and J. Tang, Energy Environ. Sci., 2015, 8, 731–759.
6. R. di Primio, B. Horsfield and M. A. Guzman-Vega, Nature, 2000, 406, 173–176.
7. N. S. Lewis and D. G. Nocera, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 15729–15735.
8. S. Solomon, G.-K. Plattner, R. Knutti and P. Friedlingstein, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 1704–1709.
9. A. Fujishima and K. Honda, Nature, 1972, 238, 37–38.
10. A. M. Alotaibi, S. Sathasivam and I. P. Parkin, RSC Adv., 2015, 5, 67944–67950.
11. A. M. Alotaibi, B. A. D. Williamson, S. Sathasivam, A. Kafizas, M. Alqahtani, C. Sotelo-Vazquez, J. Buckeridge, J. Wu, S. P. Nair, D. O. Scanlon and I. P. Parkin, ACS Appl. Mater. Interfaces, 2020, 12, 15348–15361.
12. M. Andersson, L. Österlund, S. Ljungström and A. Palmqvist, J. Phys. Chem. B, 2002, 106, 10674–10679.
13. M. P. Dare-Edwards, J. B. Goodenough, A. Hamnett and P. R. Trevellick, J. Chem. Soc., Faraday Trans. 1, 1983, 79, 2027–2041.
14. J.-W. Jang, C. Du, Y. Ye, Y. Lin, X. Yao, J. Thorne, E. Liu, G. McMahon, J. Zhu, A. Javey, J. Guo and D. Wang, Nat. Commun., 2015, 6, 7447.
15. A. Kay, I. Cesar and M. Grätzel, J. Am. Chem. Soc., 2006, 128, 15714–15721.
16. K. G. McGregor, M. Calvin and J. W. Otvos, J. Appl. Phys., 1979, 50, 369–373.
17. C. Sanchez, K. D. Sieber and G. A. Somorjai, J. Electroanal. Chem. Interfacial Electrochem., 1988, 252, 269–290.
18. R. Shinar and J. H. Kennedy, Sol. Energy Mater., 1982, 6, 323–335.
19. K. Sivula, F. L. Formal and M. Grätzel, Chem. Mater., 2009, 21, 2862–2867.
20. J. H. Kennedy and K. W. Frese, J. Electrochem. Soc., 1978, 125, 709.
21. I. Balberg and H. L. Pinch, J. Magn. Magn. Mater., 1978, 7, 12–15.
22. A. B. Murphy, P. R. F. Barnes, L. K. Randeniya, I. C. Plumb, I. E. Grey, M. D. Horne and J. A. Glasscock, Int. J. Hydrogen Energy, 2006, 31, 1999–2017.
23. J. Deng, J. Zhong, A. Pu, D. Zhang, M. Li, X. Sun and S.-T. Lee, J. Appl. Phys., 2012, 112, 084312.
24. M. Forster, R. J. Potter, Y. Yang, Y. Li and A. J. Cowan, ChemPhotoChem, 2018, 2, 183–189.
25. Y. Ling, G. Wang, D. A. Wheeler, J. Z. Zhang and Y. Li, Nano Lett., 2011, 11, 2119–2125.
26. A. Pu, J. Deng, M. Li, J. Gao, H. Zhang, Y. Hao, J. Zhong and X. Sun, J. Mater. Chem. A, 2014, 2, 2491–2497.
27. T. Wang, W. Luo, X. Wen, Z. Zou and W. Huang, ChemNanoMat, 2016, 2, 652–655.
28. K. L. Hardee and A. J. Bard, J. Electrochem. Soc., 1975, 122, 739.
29. R. K. Quinn, R. D. Nasby and R. J. Baughman, Mater. Res. Bull., 1976, 11, 1011–1017.
30. S. Shen, P. Guo, D. A. Wheeler, J. Jiang, S. A. Lindley, C. X. Kronawitter, J. Z. Zhang, L. Guo and S. S. Mao, Nanoscale, 2013, 5, 9867–9874.
31. G. Wang, Y. Ling, D. A. Wheeler, K. E. N. George, K. Horsley, C. Heske, J. Z. Zhang and Y. Li, Nano Lett., 2011, 11, 3503–3509.
32. M. Barroso, C. A. Mesa, S. R. Pendlebury, A. J. Cowan, T. Hisatomi, K. Sivula, M. Grätzel, D. R. Klug and J. R. Durrant, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 15640–15645.
33. F. L. Formal, N. Tétreault, M. Cornuz, T. Moehl, M. Grätzel and K. Sivula, Chem. Sci., 2011, 2, 737–743.
34. L. Xi, P. S. Bassi, S. Y. Chiam, W. F. Mak, P. D. Tran, J. Barber, J. S. C. Loo and L. H. Wong, Nanoscale, 2012, 4, 4430–4433.
35. V. S. Kumbhar, K. Lee, N. R. Chodankar, A. Al Ghaferi and J.-J. Shim, Int. J. Hydrogen Energy, 2025, 142, 159–167.
36. A. M. Alotaibi, E. Muayqil, N. Al Abass, M. A. Alhajji, A. A. Bubshait, N. E. Alhazmi and A. A. Almuqhim, Renewable Energy, 2024, 235, 121326.
37. C. Nomellini, A. Polo, C. A. Mesa, E. Pastor, G. Marra, I. Grigioni, M. V. Dozzi, S. Giménez and E. Selli, Improved Photoelectrochemical Performance of WO₃/BiVO₄ Heterojunction Photoanodes via WO₃ Nanostructuring, ACS Appl. Mater. Interfaces, 2023, 15, 52436–52447.
38. R. Han, W. Li, W. Pan, M. Zhu, D. Zhou and F. Li, Sci. Rep., 2014, 4, 7493.
39. Q. Wei, Z. Li, Z. Zhang and Q. Zhou, Mater. Trans., 2009, 50, 1351–1354.
40. S. Ponja, S. Sathasivam, N. Chadwick, A. Kafizas, S. M. Bawaked, A. Y. Obaid, S. Al-Thabaiti, S. N. Basahel, I. P. Parkin and C. J. Carmalt, J. Mater. Chem. A, 2013, 1, 6271–6278.
41. J. Tauc, Mater. Res. Bull., 1968, 3, 37–46.
42. A. Furlani, M. V. Russo, G. Polzonetti, K. Martin, H. H. Wang and J. R. Ferraro, Appl. Spectrosc., 1990, 44, 331–334.
43. R. P. Gupta and S. K. Sen, Phys. Rev. B: Solid State, 1974, 10, 71–77.
44. R. P. Gupta and S. K. Sen, Phys. Rev. B: Solid State, 1975, 12, 15–19.
45. A. P. Grosvenor, B. A. Kobe, M. C. Biesinger and N. S. McIntyre, Surf. Interface Anal., 2004, 36, 1564–1574.
46. N. S. McIntyre and D. G. Zetaruk, Anal. Chem., 1977, 49, 1521–1529.
47. K. Gajda-Schrantz, S. Tymen, F. Boudoire, R. Toth, D. K. Bora, W. Calvet, M. Grätzel, E. C. Constable and A. Braun, Phys. Chem. Chem. Phys., 2013, 15, 1443–1451.
48. A. P. Singh, N. Kodan, B. R. Mehta, A. Dey and S. Krishnamurthy, Mater. Res. Bull., 2016, 76, 284–291.
49. X. Guo, L. Wang and Y. Tan, Nano Energy, 2015, 16, 320–328.
50. H. Lu, S. Fang, J. Hu, B. Chen, R. Zhao, H. Li, C. M. Li and J. Ye, ACS Omega, 2020, 5, 19861–19867.
51. Y. Sun, L. Zhang, X. Zhou, Y. Ji, Y. Zhu, Q. Fu, X. Chen, X. Cheng and W. Yan, New J. Chem., 2025, 49, 10869–10878.
52. N. Dai, Y. Zhang, Z. Peng and J. Liu, J. Alloys Compd., 2025, 1018, 179294.
53. A. Memar, C. M. Phan and M. O. Tade, Int. J. Hydrogen Energy, 2015, 40, 8642–8649.
54. L. Xia, J. Bai, J. Li, Q. Zeng, X. Li and B. Zhou, Appl. Catal., B, 2016, 183, 224–230.
55. P. I. Kyesmen, N. Nombona and M. Diale, J. Alloys Compd., 2021, 863, 158724.
56. Y.-K. Hsu, Y.-C. Chen and Y.-G. Lin, ACS Appl. Mater. Interfaces, 2015, 7, 14157–14162.
57. B. Singh, S. Gautam, G. C. Behera, R. Kumar, V. Aggarwal, J. S. Tawale, R. Ganesan, S. C. Roy and S. S. Kushvaha, Nano Express, 2024, 5, 015006.
58. B. Singh, G. C. Behera, B. K. Pradhan, S. Gautam, V. Aggarwal, R. Kumar, M. S. Kumar, R. Ganesan, S. C. Roy and S. S. Kushvaha, J. Mater. Sci., 2024, 59, 15201–15220.
59. S. Shen, J. Zhou, C.-L. Dong, Y. Hu, E. N. Tseng, P. Guo, L. Guo and S. S. Mao, Sci. Rep., 2014, 4, 6627.
60. S. Shen, C. X. Kronawitter, D. A. Wheeler, P. Guo, S. A. Lindley, J. Jiang, J. Z. Zhang, L. Guo and S. S. Mao, J. Mater. Chem. A, 2013, 1, 14498–14506.
61. C. H. Bak, K. Kim, K. Jung, J.-B. Kim and J.-H. Jang, J. Mater. Chem. A, 2014, 2, 17249–17252.
62. B. Iandolo, H. Zhang, B. Wickman, I. Zorić, G. Conibeer and A. Hellman, RSC Adv., 2015, 5, 61021–61030.
63. H. Dotan, K. Sivula, M. Grätzel, A. Rothschild and S. C. Warren, Energy Environ. Sci., 2011, 4, 958–964.
64. R. Katoh, M. Murai and A. Furube, Chem. Phys. Lett., 2010, 500, 309–312.
65. Z. Huang, Y. Lin, X. Xiang, W. Rodríguez-Córdoba, K. J. McDonald, K. S. Hagen, K.-S. Choi, B. S. Brunschwig, D. G. Musaev, C. L. Hill, D. Wang and T. Lian, Energy Environ. Sci., 2012, 5, 8923–8926.
66. G. X. Pei, J. H. J. Wijten and B. M. Weckhuysen, Phys. Chem. Chem. Phys., 2018, 20, 9806–9811.
67. S. R. Pendlebury, M. Barroso, A. J. Cowan, K. Sivula, J. Tang, M. Grätzel, D. Klug and J. R. Durrant, Chem. Commun., 2011, 47, 716–718.
68. J. Tang, J. R. Durrant and D. R. Klug, J. Am. Chem. Soc., 2008, 130, 13885–13891.

---